Set micro-network to launch integrated circuit micro-letter public number: ' Every day IC ', major news immediately released, every day IC, every day set micro-net, product micro-Cheng! Replication Laoyaoic Micro Letter public number search adds attention.
1. Mainland AI chip new venture financing blowout;
In the era of PC and Smartphone, China's general-purpose chip industry is almost blank. But in the era of AI, the Chinese and foreign capital injections, a number of the mainland AI chip start-up companies are rapidly rising.
According to "Wall Street News" reports, in addition to Baidu, Huawei and other earlier engaged in chip research and development of information companies, the mainland AI chip start-ups are mostly established in 2014 to 2016, and in 2017 began to receive media and capital attention. Especially in the past few months, AI chip companies have ushered in a wave of "blowout-type" financing explosion.
First, "Cambrian" announced the completion of the $100 million in a round of financing at the end of August, followed by the "horizon" announcement last month of a A-plus round of billions of dollars in financing from Intel's Intel investment. Others include: Deep technology, Kneron, Wei-Jing Intelligence and other start-up companies, but also get a large number of financing.
Qualcomm announced last week investment in nine Chinese artificial intelligence algorithms, chips, IoT and other innovative companies, including Shenzhen AI chip start-up company Kneron (energy).
In addition, Alibaba has invested five chip companies in the past two years. Alibaba 2015 and Hangzhou Zhongtian Micro System Co., Ltd. in-depth cooperation to jointly develop the use of IoT "cloud Chip" (Yun on Chip) architecture. January 2016, Alibaba shares Zhongtian Micro become the largest shareholder.
In November 2016, Alibaba and Tencent jointly led a $23 million C-round financing of "barefoot networks", a programmable chip company. Barefoot Networks developed the world's first programmable chip, a chip called "tofino", supposedly twice times faster than all the chips currently on the market, capable of processing network packets at a rate of 6.5 megabits per second.
Coupled with the recent three months of investment in the Cambrian, deep Kam Technology and Kneron, Alibaba has been invested in five chip companies, including the focus of the Cambrian in the mobile phone business, deep Kam focus on the security field, Kneron focus on intelligent home and intelligent security.
The rapid development of the mainland's semiconductor industry has also attracted the attention of the Korean media, the Chosun Ilbo pointed out, in South Korea's leading semiconductor sector, China has also begun a strong pursuit. According to the International Association of Semiconductor Equipment and materials (SEMI), three years from this year to 2019, China plans to build a total of 15 semiconductor factories, compared with the same period in South Korea's three, Japan's four, Taiwan's seven add up to more.
Jinchang, a professor of science and technology policy at Hanyang University in South Korea, said, "Chinese semiconductors are at least able to control huge domestic markets in the short term","Korea's semiconductor industry cannot revel in its unprecedented best performance, it should have a sense of crisis that could be taken away by China to take the lead in manufacturing." 」
2. Ali department AI Chip battle team formed a heavy investment to accelerate the layout;
In the PC and mobile era, mainstream chip manufacturers are still being Intel, Qualcomm, MediaTek, Samsung and other long-term control, the advent of AI, it seems to the mainland of the chip industry to bring a new change of bureau and Opportunity, AI chip enterprise Touce momentum, even the internet giant Ali in the past three months, a continuous investment of three AI chip companies, Cumulative Ali series AI Chip network has reached five, the first embryonic. A few days ago Qualcomm Venture announced investment in 9 mainland artificial intelligence algorithms, chips, IoT and other fields of innovation companies; In addition to Qualcomm, the semiconductor giant Intel is also locked in the mainland AI and unmanned aircraft, IoT related industries active investment in the layout. Among them, the content of Qualcomm's investment in addition to the business of soup technology, motorcycle, but also include the AI chip new creation resistance (Kneron), also will be able to rise to the public's vision. It is noteworthy that the resistance to the financing of the party is not Qualcomm, Qualcomm is only as a participant, led by Alibaba founder Fund. This is also Ali or Ali related companies following the Cambrian, deep Kam Technology (by the ant gold Clothing to vote), within three months after the big investment of the third AI chip company. Internet giants grasp AI chip entry point Overview of these major AI chip company's growth history, most of which was established in 2014-2016 years, the 2017 AI applications are widely watched by the market, become the mainland AI chip floating on the stage of the rise of the year, and more access to the market a lot of exposure and capital side of the high And the funding for each family is small. At the same time Ali and Huawei, Tencent, Baidu, Lenovo and other companies are also in a variety of ways to quickly close to the AI chip industry, is because, a variety of terminal market in the vertical application of the scene, including mobile phones, smart homes, security, automatic driving and other fields, all need a variety of dedicated AI chip as the underlying architecture. The early placement of AI chips is important both in terms of market prospects and in the upstream resources of the AI industry chain. Second, the AI chip in the smartphone terminal application of the fiery, each AI chip in the future at AI entrance, such as intelligent sound, robot in the core strategic position, prompting the internet giants to realize the need to accelerate the layout of AI chip development. The recent cloud Habitat conference, Alibaba set up the ' Dharma Institute ', Ma Yun announced the launch of the ' new technology Strategy ' program, machine learning, chip, networking, operating systems, biometric core technology, the formation of a new team, the establishment of novel mechanisms and methods. Among them, Alibaba CTO Zhang Jianfeng more mention, Ali Research and development of the ' Ali Ai Core ' will soon be available, which will promote Aliyun services more stable, controllable and efficient; And AI chip has become a strategic weapon, as the giant's Ali natural will not let go. On the other hand, from the chip industry, AI chip to build there are two of challenges: one is to do the chip of the capital human investment is larger, the second is to have enough scenes to be used. These two points are extremely scarce for startups, so it is important to ' backer ' through the introduction of capital selection. In addition to its sophisticated cloud computing platform, Ali has a variety of application scenarios, as well as powerful new retail and online sales resources. In particular, Aliyun generates a huge amount of user data, which is used to train the most fertile nutrients in the AI scene, but also requires higher computational power, lower power consumption, and smaller-volume AI chips to meet operational requirements. Ali Department has accumulated investment of five AI Chip enterprises 2015 years, Ali and Hangzhou Zhongtian Micro-depth cooperation, object-oriented networking in various segments of the development of Cloud Chip (Yun on Chip) architecture. January 2016, Ali official shares Zhongtian Micro became its largest shareholder, it is reported that the proportion of nearly 30%. So far the transit micro-cumulative SOC shipments have more than 600 million. In November 2016, Ali also with Tencent led the FPGA programmable design chip public barefoot Networks $23 million C-round financing. Barefoot Networks developed programmable Chip Tofino, which is said to be twice times faster than any other chip on the market now, capable of processing network packets at a rate of 6.5 megabits per second. Plus, nearly three months of investment in the Cambrian, deep Kam Technology and Kneron, Ali system has been in the past two years, the cumulative cast of 5 chip companies, is silently layout ai chip industry, master the independent AI core chip. However, Ali invests in the company's product application scenario is not exactly the same: Hangzhou Zhongtian micro-focus on IoT, such as Low-power and information security field; The Cambrian focus in the mobile phone business and high computing needs of the data center; Deep focus on security applications focused on the field, while Resistance (Kneron) focus on smart home and smart security applications. In addition, unlike the Cambrian, deep reference technology, horizon, etc. born in mainland China, Resistance (Kneron) was founded in 2014 in the United States San Diego, 2015 years before the establishment of a branch office in Shenzhen. Founder Liu Junshing has been in Samsung, Qualcomm engaged in research and development work, because the optimistic AI so venture into the neural network depth learning NPU processor chip. Digitimes
3. Business and Soup Technology program IPO and set up research and Development center in the United States;
According to Reuters, China artificial intelligence start-up company Tom Technology founder Tang told Reuters in an interview that the company is planning an initial public offering, and is preparing to set up a research and development center in the United States as early as next year.
Earlier this month, Shang-Tang technology had just received a strategic infusion from Qualcomm, a US company. The company and Qualcomm announced last month that it will launch strategic cooperation in the field of artificial intelligence, and that the algorithm developed by the business soup technology will be applied to smart devices.
In July this year, Shang Tang technology announced the completion of the 410 million-dollar round of financing, a global artificial intelligence sector single round of the highest record of financing, Shang Tang technology at the time also became the world's highest level of financing of artificial intelligence unicorn enterprise.
Shang Tang Technology was founded in 2014, the main research direction for facial recognition, video analysis, unmanned intelligent areas such as artificial intelligence, but also a number of domestic enterprises service providers. The great heat of artificial intelligence in the 2017 became a source of confidence in business-Tang technology.
When asked about the company's development, Tang said: "Our goal is definitely not to create a small company that has been acquired, but rather a platform-oriented company like Google and Facebook, which has the original core technology."
4. National strategy cooperation ' six runway ' tackling AI;
Recently, China announced the launch of a new generation of artificial intelligence development plan, will take 13 years, will be the world's main artificial intelligence innovation Center.
How does the new generation of AI develop? Who will lead? How to break through? The Reporter interviews authoritative department, the exclusive present development plan.
15 ministries join forces to build national pool of experts
2016 Global science and technology giant AI investment has reached 30 billion U.S. dollars, the importance of artificial intelligence has become a global consensus. Compared with the world's leading countries, China's artificial intelligence in some areas of core technology to achieve an important breakthrough, but still lack of significant original results, in the basic theory, core algorithms and key equipment, high-end chip and other aspects of the big gap.
How does the new generation of AI develop? "The development of a new generation of AI is a strategic cooperation at the national level," said Wan Gang, Minister of Science and Technology. '
-Ministries formed resultant force. By the Ministry of Science and Technology, development and Reform Commission, Ministry of Finance, Ministry of Education, Ministry of Industry, CAS, Engineering Institute, Military commission of Science and Technology, China Association of Sciences and other 15 departments of the new generation of AI, the promotion of the office came into being, to promote the project, base, talent overall
At the same time, to build national experts library. By Pan Academician Ren, including Chen Chun Academician, Li Wei Academician, Gawain academician and other 27 top experts, including the new generation of artificial Intelligence Strategy Advisory Committee was also announced.
' Gather collective wisdom from a forward-looking perspective, conduct substantive, potential research on the direction of some artificial intelligence, major issues of professionalism, and provide further consultation and evaluation for the major decision-making of artificial intelligence at the national level. ' Academician of CAs, Director of Strategic Advisory Committee Pan said.
--Two kinds of attributes are highly fused. ' Grasp the characteristics of the high integration of the attributes of AI technology and social attributes, and give full play to the government's important role in planning, guiding and policy support. Wan Wan told a story: ' Many foreigners are curious to ask me, how did you get the fourth generation of mobile communication? They felt very confused. I said very simple, from the beginning of the expo we are doing pilot, and then in Hangzhou, Guangzhou to do experimental area. And so experience slowly mature, and then engage in dozens of of cities, the final experiment area almost covers all areas, not so solved. '
--Open up six ' runways '. The new generation of artificial intelligence development plan, the establishment of the open and cooperative artificial Intelligence technology innovation system, the construction of a safe and convenient intelligent society, the construction of a safe and efficient intelligent infrastructure system, strengthen the artificial intelligence field of civil-military integration, cultivate high-end high efficiency intelligent economy, forward-looking layout of a new generation of
Pan said that previously only the technical level of a runway, the new generation of artificial intelligence Development program in the original runway has increased five of runways, expanding the development of artificial intelligence space.
Qin Yong, Director of Science and Technology Division, said the new generation of major science and technology projects to choose Large data intelligence, Cross-media Intelligence, group intelligence, hybrid enhanced intelligence and autonomous intelligence system five main direction, from the basic theory, support system, key technology, innovation and application of four aspects of systematic, forward-looking task layout.
Identify the first 4 national new generation artificial intelligence open innovation platform
After full investigation and demonstration, the first batch of national new generation artificial Intelligence open innovation platform is determined: Relying on Baidu, Aliyun, Tencent, Hkust, the construction of automatic driving, urban brain, medical imaging, intelligent voice 4 national new generation of artificial intelligence open innovation platform.
' These four areas, our country has been at the forefront of the world. ' Chih, Deputy Secretary of the Ministry of Science and Technology, said, accelerating the implementation of a new generation of major AI projects, as soon as possible in the basic frontier field of breakthrough. The open and innovative platform will play a greater role in bringing together innovative resources and promoting public sharing.
6 years ago, Hkust launched the world's first voice cloud open platform. At present, the daily processing frequency of the platform reaches 4 billion times. In speech synthesis, China is an international leader. ' said Hu Yu, vice president of the Hkust.
In recent years, HKUST has also launched innovative cooperation with the security system. The test conducted in Anhui in 2016 showed that the accuracy rate of artificial intelligence was very high. AI also acts as a court clerk, and after testing, intelligent speech recognition participated in the trial to achieve practical results, the length of the trial shortened by 30%.
' Let innovation produce real influence and create real value. ' The goal of the city's brain is to build a shared platform that is based on artificial intelligence, big data, cloud computing, and an open computing data application layer to build an open future ecosystem for urban administration, public services and other industries ' smart applications, said Huaxian, Alibaba's vice president.
Li Meng, Deputy Minister of Science and Technology, said the deadline for the new generation of major science and technology projects to be implemented by 2030 was only the first, and the future will continue to expand according to the needs of planning and project implementation.
The new focus of international competition is the new competition in depth learning, cross-border integration, human-computer synergy, and group intelligence. Wan said.
Sharing results in the collision to achieve innovation
The development of new generation of artificial intelligence needs to be made breakthroughs. Experts believe that the group intelligence Open will become a new generation of artificial intelligence development characteristics.
This July, Baidu officially opened the ' Apollo ' automatic driving platform, partners now more than 70, from all over the world. The Apollo program aims to provide an open, complete, secure, open and innovative platform for partners in the automotive and self-driving sectors, bringing together the power of many participants to promote the development of automatic driving technologies.
It is an open source system and the world's largest automatic driving ecology. Baidu Company vice president Shan said, Baidu will invest 10 billion yuan in the future, to support 100 start-up companies, to create the world's advanced level of automatic driving system open innovation platform.
The reporter learned that the Tencent Artificial Intelligence Laboratory last year, the establishment of a comprehensive basic research and rapid technical applications together, gathered more than 70 domestic and foreign first-class scientists and more than 300 experienced engineers.
Tencent's AI medical products Tencent found that the accuracy of early esophageal cancer increased significantly. Liu Yongxhen, general manager of AI Laboratories, a Tencent company, said that corporate sharing of research and application results will rapidly advance technology iterations, allowing traditional industries such as healthcare to generate new kinetic energy.
' To enhance open source openness and policy guidance, and to build an open and collaborative artificial intelligence technology innovation new body. Wan Wan said: ' In the past we used to say that the experiment was a thought-and-inspiration, now different, need to turn the thinking into a challege, in the course of the impact of academic exchanges to stimulate innovation. '
Xinhua News Agency, Beijing, November 23
Baidu
Autopilot
This July, Baidu officially opened the ' Apollo ' automatic driving platform, partners now more than 70, from all over the world. The Apollo program aims to provide an open, complete, secure, open and innovative platform for partners in the automotive and self-driving sectors, bringing together the power of many participants to promote the development of automatic driving technologies. It is an open source system and the world's largest automatic driving ecology.
Tencent
Medical Imaging
Founded last year, the Tencent Artificial Intelligence Laboratory combines comprehensive basic research with rapid technical applications, gathering more than 70 leading scientists at home and abroad and more than 300 experienced engineers. Tencent's AI medical products ' Tencent looking for a shadow ', found that the accuracy of early esophageal cancer increased significantly. Enterprise sharing research and application results will rapidly advance the technology iteration, so that traditional industries such as medical treatment can produce new kinetic energy.
Aliyun
City Brain
The goal of the city's brain is to build a shared platform of major open computing data applications based on artificial intelligence, large data, and cloud computing, and to build an open future ecosystem for urban administration, public services, and other intelligent applications.
Hkust News Flight
Intelligent Voice
6 years ago, Hkust launched the world's first voice cloud open platform. At present, the daily processing frequency of the platform reaches 4 billion times. China is already a global leader in speech synthesis. In recent years, HKUST has also launched innovative cooperation with the security system. Nanfang Daily
5.IEEE released three new artificial intelligence standards related to high-level ethical issues;
IEEE (American Institute of Electrical and Electronics Engineers) has recently announced three new standards for the development of artificial intelligence. IEEE is known as the world's largest advanced technology professional organization. According to a recent report, their AI standards will give priority to humans because they promise to keep pace with the development of the AI field.
In the published "Ethical Design: The vision of human welfare and artificial intelligence and autonomous systems priorities", a total of eight sections describe the new AI development issues. are: general principles, artificial intelligence system assigned value; To guide the methodology of ethics research and design; The safety and welfare of general artificial intelligence and super artificial intelligence; Personal data and personal access control; Reconstruct the automatic weapon system; (a) economic/humanitarian issues; Legal.
General principles involve high-level ethical issues that apply to all types of artificial intelligence and autonomous systems. In the determination of general principles, the main consideration is three factors: the embodiment of human rights; Priority is given to maximizing the benefits to mankind and the natural environment; Weaken the risk and negative effects of artificial intelligence.
The principle of human interest requires consideration of how to ensure that AI does not violate human rights. The principle of responsibility concerns how to ensure that AI can be accountable. In order to solve the problem of fault, avoid public confusion, AI system must be at the procedural level of accountability, proving why it operates in a certain way. The principle of transparency means that the functioning of an autonomous system must be transparent. AI is transparent, meaning people can find out how and why they make specific decisions.
In the context of how to embed human norms and ethical values in AI systems, the report says it is critical that AI systems become more autonomous in making decisions, manipulating their environment, and allowing them to adopt, learn and comply with the norms and values of the societies and groups to which they serve. The goal of embedding value in AI systems can be achieved in three steps:
First, to identify the norms and values of a particular society or group;
Second, these specifications and values are programmed into AI systems;
Third, evaluate the effectiveness of the specification and value of the AI system, i.e. whether it is compatible with the actual norms and values.
Although the above-mentioned studies have been ongoing, these fields include machine ethics (Machine morality), Machine ethics (Machine Ethics), moral machinery (moral Machine), and the value of the same theory (values Alignment), Artificial morality (Artificial morality), security Ai, friendly AI, etc., but developing a computer system that recognizes and understands human norms and values, and allows them to consider these issues when making decisions, has been troubling people. There are currently two main paths: the Top-down path and the bottom-up path. Research in this area needs to be strengthened.
The report also points to the need to guide the methodology of ethical research and design, the safety and well-being of general artificial intelligence and super artificial intelligence, the restructuring of autonomous weapons systems, and the problems of eight of artificial intelligence developments such as economic/humanitarian issues. The report collects more than 100 ideas from academic, scientific, and government-related departments in the field of artificial intelligence, integrating expertise in a wide range of areas such as artificial intelligence, ethics, philosophy and policy. Its revised version of the EAD (ethically aligned design) will be available by the end of 2017, with a report added to 13 chapters, with the best of more than 250 global thought leaders sharing.
Satoshi Tadokoro, chairman of the IEEE Robotics and Automation Association, explains why they want to set the standard: ' Robots and automated systems will bring great innovation to society. More recently, the public has become increasingly concerned about possible social problems and the potentially enormous potential benefits. Unfortunately, in these discussions, there may be some error messages from fiction and imagination. '
Tadokoro continued: ' IEEE will introduce knowledge and wisdom based on recognized facts of science and technology to help achieve public decision-making, maximizing the overall interests of humankind. ' In addition to AI ethical standards, there are two of other AI standards being introduced into the report, each headed by a field expert.
The first criterion is: ' The ethical driving standard of a robotic system, an intelligent system and an automated system '. This standard explores ' push ', and in the world of artificial intelligence, it refers to the subtle actions that AI affects human behavior.
The second standard is ' automatic and semi-automatic system failsafe design standards '. It contains automated technologies that may cause harm to humans if they fail. For now, the most obvious problem is self-driving cars.
The third criterion is the ' ethical measurement of artificial intelligence and the well-being of automated systems '. It illustrates how advances in AI technology benefit human beings.
These standards may be implemented earlier than we thought, because companies like OpenAI and DeepMind are increasingly advancing AI development, and even creating artificial intelligence systems that can learn and expand their ' intelligence ' domains. Experts believe such artificial intelligence could destabilize the world, leading to massive unemployment and war, and even to the creation of ' killing weapons '. Recent important discussions at the United Nations have prompted serious reflection on the need for stricter regulation of the use of artificial intelligence as a weapon. Lei Feng Net
6. Musk: The success rate of preventing AI from destroying humans is only 10%.
The Ellon Masc, known as the Iron Man, always likes to make predictions, you may have seen some examples, such as we live in a computer simulation world, or by 2025, Tesla's market value will surpass Apple, or humans can make Mars habitable by nuclear weapons.
Musk recently asked another question: Artificial intelligence could kill all of us. In a recent conversation, Musk said that all efforts to secure artificial intelligence were ' 5% to 10% success rates '. Over the past few years, Musk has repeatedly warned of the threat of artificial intelligence. As early as 2014, he predicted that Ai was on the verge of ' serious danger '.
Just last year, Musk announced that humans had actually lost the battle with artificial intelligence, and that the only way to defeat them was to join them. "We will be far behind in any progress in the field of artificial intelligence," Musk says. The benign development of super intelligent AI is that our IQ is far lower than that of AI, we are like pets, or home cats. '
To solve this problem, Musk became an important supporter of neural lace, a reticular implant containing electrodes that can monitor the function of the brain. Essentially, it's a ribbon cable between your brain and your PC, allowing simultaneous data transfer. This is the foundation of Neuralink, Neuralink is one of Musk's most recently established start-up companies.
Maybe it's time to ring the bell, even if it's just a little. As the leader of Neuralink and OpenAI, Musk recently developed an artificial intelligence technology that can teach itself and invests heavily in the field of artificial intelligence. Wei Feng Network